The complimentary strengths of the AmpereOne processor and NVIDIA GPUs for AI development are available on NextComputing’s Edge XTP tower workstations. These NextComputing systems are well suited for a variety of AI use cases, from early-stage research and model development to deployment in dynamic, real-world environments.
Edge XTP is configured for your specific use case, with support for the latest and most powerful GPUs and CPUs as well as high capacity for RAM and storage. The high performance and data throughput of Edge XTP workstations make them ideal tools to leverage the power of large language models (LLMs) for creating new content with generative AI or to training agentic AI for autonomous computing.
View tower workstations featuring Ampere CPUs
AmpereOne Processors
AmpereOne processors feature outstanding performance and power efficiency. They offer linearly scalable throughput, making them ideal as a machine learning tool. They are perfectly suited for data-intensive tasks such as the development of AI-based cloud services (recommender engines, predictive analytics, natural language processing, and computer vision applications).
- High Core Count and Parallelism: AmpereOne processors can have up to 192 cores, providing massive parallelism for AI inference workloads like natural language processing, image recognition, and generative AI. This architecture is designed to handle high-density computing and can deliver consistent, low-latency performance.
- Energy Efficiency: Ampere’s architecture is known for its high performance per watt, which is crucial for managing the operational costs and environmental impact of running AI workloads continuously in cloud environments.
- Scalability: AmpereOne is built for dense containerized deployments. Applications scale linearly as additional single-threaded cores are utilized. This makes AmpereOne an ideal platform for cloud native applications sensitive to latency variability or needing strict SLAs. Services integrated with AI ranging from visual analytics and natural language processing to classical machine learning pipelines are workloads that benefit from this level of consistency, elasticity and determinism.
- Incredible Speed: With 128 lanes of PCIe Gen5, AmpereOne can support up to eight x16 devices attached to a single socket. This provides a high-speed interface for data transfer to and from the GPU.
- Built-In AI Inference: Ampere helps customers achieve superior performance for AI workloads by integrating optimized inference layers directly into popular AI frameworks like PyTorch, TensorFlow, and ONNX Runtime. This allows for seamless deployment without requiring code changes or model conversions.
These features lead to highly predictable performance. More compute intensive tasks can be performed with fewer systems, lowering TCO. AmpereOne architecture drives faster, more predictable AI services that scale seamlessly with demand, helping data centers maximize efficiency and reduce infrastructure costs.
The Latest NVIDIA GPUs for AI
NVIDIA GPUs are versatile and powerful tools for AI development, striking a balance between high-end performance and cost-effectiveness. They are particularly well-suited for a wide range of AI workloads, from training and inference to graphics-intensive applications.
NVIDIA L4 and NVIDIA L40 GPUs
- AI and Deep Learning Capabilities: The L4 and L40 GPUs are built on the Ada Lovelace architecture and feature 4th-generation Tensor Cores, which are specifically designed to accelerate AI training and inference.
- Large Memory Capacity: With 48 GB of GDDR6 memory, the L40 can handle very large datasets and complex models without memory bottlenecks. This is especially important for training large language models and other data-intensive tasks.
- Scalable, Efficient Performance: This architecture delivers groundbreaking performance for deep learning and inference applications, and it is optimized for 24/7 enterprise data center operations, ensuring reliability and uptime while maintaining energy efficiency and lower total cost of ownership.
NVIDIA RTX PRO™ 4500, 6000 Blackwell GPU
Designed for professionals who demand the best, NVIDIA RTX PRO Blackwell is the most powerful professional RTX GPU. It delivers next-level AI and neural rendering with 5th Gen Tensor Cores, 4th Gen RT Cores, and Blackwell’s advanced CUDA architecture, accelerating data workflows like never before.
NVIDIA H100, H200 NVL Tensor Core GPUs
The NVIDIA H100 and H200 NVL are both cutting-edge data center GPUs built on the Hopper architecture. The H100 is a powerful, well-established solution with 80 GB of HBM3 memory and 3.35 TB/s of bandwidth, while the H200 NVL nearly doubles the memory capacity to 141 GB of faster HBM3e memory. These cards particularly effective at accelerating inference for massive large language models and other memory-bound tasks, offering faster processing, lower latency, and improved energy efficiency per token, thereby lowering the total cost of ownership for large-scale enterprise AI deployments.
The NextComputing Solution
Edge XTP workstations combine the AmpereOne’s high core count and energy efficiency with the NVIDIA GPU’s specialized AI acceleration and large memory capacity to create a powerful and balanced system. The AmpereOne can efficiently handle the general-purpose computing and data management, while the GPU can accelerate the computationally intensive parts of AI and deep learning. This synergy allows for high-throughput AI inference and training, making it an ideal platform for a wide range of AI development tasks.